Séminaire Lab-STICC le 05/11/10 : Audio blind source separation: robustness and sparsity
Dans le cadre des séminaires LabSTICC de recherche en signal, images et communications, Si Mohamed AZIZ SBAI, doctorant au département SC de Télécom Bretagne, présentera ses travaux intitulés "Audio blind source separation: robustness and sparsity".
Date, heure et lieu :
Vendredi 5 novembre à 14h (durée : 1h environ), en salle K02-109A de Télécom Bretagne.
Résumé :
In a sparsity framework, we address the problem of blind source separation in the underdetermined and instantaneous mixture case. The proposed method is based on an algorithm developed by Aïssa-El-Bey and al. This algorithm requires a good choice of the noise threshold but does not take into account the noise contribution in the inversion process. In order to overcome these drawbacks, this presentation presents a robust underdetermined blind source separation approach. Robustness is achieved by estimating the noise standard deviation and using this estimate in the inversion process and the expression of the noise threshold. The good performance of the proposed method is shown by comparison with state-of-the-art methods.
Keywords: Sparsity, robustness, blind source separation, noise variance estimation.
Séminaire Lab-STICC le 18/11/10 : Variational Methods for Missing Data Interpolation In Geophysical Image Sequences
Dans le cadre des séminaires LabSTICC de recherche en signal, images et communications, Sileye Ba, postdoctorant au département SC de Télécom Bretagne, présentera ses travaux intitulés "Variational Methods for Missing Data Interpolation In Geophysical Image Sequences".
Date, heure et lieu :
Jeudi 18 novembre à 14h (durée : 1h environ), en salle K01-232.
Résumé :
One of the goal of The CREATE project is the characterization of the dynamics of extreme oceanic events. The extreme event we focused on are the oceanic fronts which are narrow zone separating water masses that support the high gradients geophysical data such as the sea surface temperature. Studying front dynamics requires their detection and tracking in the relevant geophysical data. However, because of the cloud coverage, geophysical satellite images acquired in the visible range are produced with missing data. In this talk, we present variational methods for missing data interpolation in static images and dynamic image sequences. The first part of the talk will mainly focus on the missing data interpolation in a single type of geophysical data. In the second part of the talk, we will address the joint missing data interpolation problem for different type of geophysical data. This will allow the use of joint relationship between these data as a priori information.
CS 83818 - 29238 Brest Cedex 3
Tél. : +33 (0)2 29 00 11 11
Fax : +33 (0)2 29 00 10 00
Les membres :
